nTop Live: FE Simulation of Gyroid Heat Exchangers from nTopology to ANSYS

[@mansourehrezapourian1404]

your training was super good. thank you so much

[@madladdan]

Are there any videos that explain monitoring, convergence and optimizing the heat exchanger?

[@최진범-y7r]

How did you share topology between the fluid domain and the solid?

[@ococcm]

so im facing a problem whem i import the .scdoc files from spaceclaim into Fluent! i keep getting this following error: Error: FMD conversion of "C:/Users/majd_/Desktop/BA/Ansys/Ansys 8/Ba_files/dp0/Geom/DM/Geom.scdoc" has failed!Error Object: #f

[@jaburelll]

Hello, I was facing the same problem but I found one solution. I might be wrong, but it worked for me, so I guess Fluent mesher natively uses .fmd format and when importing a .scdoc file to Fluent mesher, it converts it back to .fmd to start the meshing process. For me, the conversion was failing as well because that error is encountered when you create named selections (NS's) in SpaceClaim and something goes wrong when converting and it can't correctly translate NS's from .scdoc to .fmd. The procedure in this video is that the faces of the cells are marked by selecting them using single selection tool (double click a face and increase the slider that appears). Basically, I think this selection tool is based on the curvature of the face that you are selecting first so when you increase that slider, you increase the curvature so it selects more faces with higher curvature (or angle btw faces, I'm not 100% sure so correct me if I'm wrong) so you can capture more of the faces within the core of the HEX. Though, selecting more and more faces and trying to accurately capture all the faces of the HEX core, you can select the same faces (or edges) in 2 or more NS's, therefore conflicting Fluent .fmd conversion because you have at least one face (or edge) in two different NS's. Basically you have to be very careful when creating NS's and for me it was a very trial and error process and took a lot of tries. You basically want to capture all of the faces/edges, but it's hard to do so because in theory TPMS have infinitely curvature so you better not increase that slider that much and do what Mikey did in this video at 6:07 and leave some faces unselected, but you can use Ctrl+click to capture the ones missing if you can't do it from first try (act never could so had to multiple select many times holding Ctrl). One thing you can do is try and save as .fmd and see if it works or you might get an error (in SpaceClaim) telling you that one or more faces are part of multiple NS's. And then just go normally about the process, but instead of an .scdoc file you use .fmd for Fluent, because when saving .scdoc you might not get the error in SpaceClaim, but you get it in Fluent as you're importing the file (the error in your comment), and this is time consuming because you'd think you'd have got it right in SpaceClaim but when starting Fluent you got it wrong. What you can also do is saving intermediate .scdoc files when creating working pairs of NS's (Cold_Solid_intf and Cold_Fluid_intf). First, try saving as .fmd, see if it works, if it does, save it as intermediate .scdoc file, then proceed onto creating the next pair of NS's and if the error occurs, then go back, import the .scdoc file with the already working NS's up to that point, and try do it again and again. Furthermore, if you miss selecting some faces, Fluent Mesher doesn't leave those faces undealt with, they will appear in 'Update Boundaries' tab as 'facets.1', facets.2', etc leave those as walls (you can hover over them and will be highlighted in yellow but those are not exactly the ones translated into Fluent Solver so don't mind about them at that point). Fluent Solver then automatically translates and creates slightly different NS's for them (as wall BC's). Most of the times those missed faces within the core because you probably just couldn't see them or select them are created in independent NS's (you have to display them and see which one is which and where they're located just by visual inspection or checking which zone is adjacient to). This is fortunate enough that you don't have to worry about them in Fluent Mesher because in Fluent Solver you can use the 'merge zones' option to merge them with the rest of the core NS's from which they should've been part of (see adjacient zones). Hope this helps clarifying some things, though being a bit late, or helps anyone else getting over this problem. God bless.

[@madladdan]

Did you fix this?

[@ococcm]

I don’t remember to be honest. I took another approach.

[@madladdan]

@@ococcm Yea I ignored fluent and stuck with discovery

[@roopanr4302]

How to create baffles in ntop tpms structure